Content uploaded by Jana Bressem
Author content
All content in this area was uploaded by Jana Bressem on Mar 26, 2015
Content may be subject to copyright.
Semiotica 184–1/4 (2011), 53–91 0037–1998/11/0184–0053
DOI 10.1515/semi.2011.022 © Walter de Gruyter
Rethinking gesture phases: Articulatory
features of gestural movement?*
JANA BRESSEM and SILVA H. LADEWIG
Abstract
This paper presents a proposal for the description of gesture phases derived
from articulatory characteristics observable in their execution. Based on the
results of an explorative study examining the execution of gesture phases of ten
German speakers, the paper presents two sets of articulatory features, i.e.,
distinctive and additional features by which gesture phases are characterized
from a context-independent and context-sensitive point of view. It will be shown
that gesture phases show a particular distribution of the features, thus distin-
guishing one phase from another. Furthermore, changes in the execution of
phases in linear successions can be described by means of features. Contrary
to other accounts, whose focus on gesture phases is primarily in relation to
speech and/or adjacent phases, this proposal concentrates on the visible phys-
ical characteristics of gesture phases.
Keywords: gesture phases; articulatory features; form-based analysis;
c ontext-independent and context-sensitive description; linguistic
perspective
I think it would be worthwhile to pursue a pro-
gram of research on the perception of action to
try to identify what appear to be the movement
features that people rely upon to separate “ges-
ture” (actions perceived as produced to “say
something,” etc.) from other kinds of actions.
— Kendon (1996: 8)
When watching people talk, it can be observed that speakers quite frequently
move their hands. Speakers might do this because they want to describe what
a particular object looks like. They might depict its form, size or shape. At
Bereitgestellt von | TU Chemnitz
Angemeldet
Heruntergeladen am | 26.03.15 10:39
54 J. Bressem and S. H. Ladewig
other times, speakers use their hands to point at things in their surroundings,
such as objects in the real world or the person they are talking to. Also, s peakers
frequently use their hands to convey their attitude to what is being talked about.
They might mark the utterance as a question, a negation, a doubt or to belittle
another person.
When doing so, speakers are engaging in the act of gesturing, i.e., they per-
form “communicative movements of the hands and arms, which, similar to
language are used to express the thoughts, feelings, and intentions of a speaker
. . .” (Müller 1998: 13, our translation). When people gesture, they move their
hands and arms in a particular succession. Starting from a relaxed position,
such as on their lap or on a table, they move them to a place in front of their
body, where they may perform further movements, and then back to a relaxed
position again (see Figure 1). These successions, rst dened by Kendon
(1980), are referred to as “gesture phases” and describe the different movement
phases observable in the execution of gestures.
The meaningful part of the gesture1 — the part people rely on in their inter-
pretation of a gesture — is the stroke (Kendon 1980). In order to perform a
stroke, the hands need to prepare for its execution during the phase referred to
as preparation (Kendon 1980). The stroke may be followed by a retraction, a
phase in which the hands relax and move back into a rest position. These ges-
ture phases build higher-level units, namely: 1) the gesticular phrase/gesture
phrase, which is composed of a preparation and stroke and 2) the gesture unit,
which is “demarcated as extending from the moment p begins the excursion of
the limb to the moment when the limb is in rest again” (Kendon 1980: 212) and
may include one more gesture phrases.2
Kendon’s pioneering work on the description of gesture phases, along with
the work of others that have followed, has been essential to the study of co-
verbal gestures. It has shown that gestural movement sequences can be broken
down into a succession of different phases, which correspond to units at speech
level (Kendon 1972a). Furthermore it has shown that gestures form larger
units, which match higher-level units at the verbal level (Kendon 1972a). More
importantly however, it has provided a technique for detailed accounts of ges-
Figure 1. Succession of gesture phases
Bereitgestellt von | TU Chemnitz
Angemeldet
Heruntergeladen am | 26.03.15 10:39
Rethinking gesture phases: Articulatory features of gestural movement? 55
tures and their relation to speech, and has proven the fact that “speech and
movement appear together as manifestations of the same process of utterance”
(Kendon 1980: 208).
It has been more than thirty years since Kendon’s rst account of the struc-
ture of gesture phases was published. Since then, only minor changes have
been made (see Section 2). To this day, the relation of gesture to speech is
mainly of interest to prove the close relationship between speech and gesture
(see for example Bohle 2007; Kita, van Gijn, and van der Hulst 1998; Kendon
2004). Questions on the characteristics and features of different gesture phases,
or questions on the status of gesture phases in the coding and in the analytical
process are rarely addressed. Furthermore, questions on the role of gesture
phases for the structure and composition of the medium of “gesture” itself, as,
for instance, in the building of higher-level units (Kendon 1972a, 1980, 2004)
or their recursive embeddings within these units (Fricke 2008) — properties
that have so far been ascribed to speech only — have been paid attention to
only selectively.
This paper therefore concentrates on the phases themselves and tries to in-
spire a linguistic perspective on the description of gesture phases. We will in-
troduce them as minimal units of analysis, which can be described on their own
and in relation to each other. In doing so, we approach gesture phases from the
perspective of gesture analysts. We try to explicate features that an analyst
perceives when coding gesture phases and relies on when describing their
forms and functions. Thus, the questions we want to answer in the paper are:
What do researchers perceive and attend to when segmenting and coding ges-
ture phases? Are there features that aid not only the segmentation and coding
process, but can also be used for a form-based characterization of the various
gestures phases?
The aim of this paper is therefore to spell out the articulatory features at-
tended to, systematize them, and classify gesture phases according to them.
The proposal is not meant to be a guide for gesture phase coding. We challenge
trained gesture analyst’s perspectives on the object of investigation and their
implicit use of these features in the whole coding process. We do this in order
to provide a step in the direction of gesture phase coding only on the basis of
articulatory features.
For the approach to be presented, we therefore conceive gesture phases as
potentially separable units of analysis. Each gesture phase can be investigated
and described on its own, independent of its integration in the ow of move-
ment of a gesture unit (Kendon 2004). This analytical separation is assumed
for all gesture phases and is a starting point for our articulatory description.
So far, investigations in gesture studies only implicitly conceive gesture
phases as separable units of analysis and, moreover, mostly with respect to the
stroke. Thus, particularities in the execution of gesture phases and possible
Bereitgestellt von | TU Chemnitz
Angemeldet
Heruntergeladen am | 26.03.15 10:39
56 J. Bressem and S. H. Ladewig
characteristic features have mostly been related to the stroke. Kendon for ex-
ample writes, “the stroke is the phase of the excursion in which the movement
dynamics of “effort” and “shape” are manifested with greatest clarity” (2004:
112). Similarly, Sowa points out that “it is assumed that the overall effort
d uring a stroke is comparably high, while the other dynamic phase types are
marked by a more or less constant increase or decrease from or to a low effort
value” (2006: 195). Due to its special characteristics, the stroke has been an
object of investigation from very early on in gesture research. Several studies
have concentrated on the synchronization of the stroke with respect to the co-
occurring speech and in particular to the correlation of the stroke and accents
in speech (Efron 1972 [1941]; Kendon 1972a; Condon and Ogston 1967; Mc-
Clave 1991; Nobe 2000; Loehr 2006; Queck et al. 2002; Yassinek et al. 2004).
Other analyses have shown that a gesture waits for the speech in that the stroke
for example might be “suspended” (Seyfeddinipur 2006) in order to be syn-
chronized with the verbal component and to a create a “gesture-speech en-
semble” (Kendon 2004: 127; see also Kita, van Gijn, and van der Huslt 1998
with respect to pre-stroke and post-stroke holds and their relation to the speech
production process).
Apart from the stroke or hold, particular analyses have also detached other
gesture phases from the ow of movement and investigated their relation to
co-occurring speech. Seyfeddinipur (2006), for instance, was able to show that
gestures can be interrupted in their execution. “Specically, in disuent utter-
ances, gesture suspension tended to occur within a preparation or a stroke or
right after a preparation” (Seyfeddinipur 2006: 143) thus leading to changes in
the temporal organization of the gesture phrase (Kendon 2004).
Examples referring to particular gesture phases independent from their con-
nection to the remaining gesture phases can also be found in specic types of
gestures phases as “interrupted preparation/stroke” (Seyfeddinipur 2006: 109)
and “partial retraction” (Kendon 1980; Seyfeddinipur 2006), for instance.
The status of gesture phases as units of description and the assumption that
gesture phases can be treated on their own, apart from their integration in a
gesture unit, have only rarely been addressed in gesture research. Explicit ref-
erence to single phases and their relation to the utterance process yet show that
certain gesture phases are also conceived as single units of investigations. Fur-
thermore, it indicates that phases possess particular features that allow for their
recognition within the ow of movement and reect the functions they fulll
for adjacent phases and for speech.
Some attempts at describing possible features of gesture phases have been
made, especially in the eld of research on automated gesture segmentation,
(cf. Chafai, Pelauchad, and Pelé 2006; Harling and Edwards 1997; Latoschik
2000; Kahol, Tripathi, and Panchanathan 2004; Martell and Kroll 2007; Wil-
son, Bobick, and Cassell 1997). Yet none of these approaches has pursued a
Bereitgestellt von | TU Chemnitz
Angemeldet
Heruntergeladen am | 26.03.15 10:39
Rethinking gesture phases: Articulatory features of gestural movement? 57
linguistic, feature-based perspective that relates the different phases and their
articulatory features to each other, on the level of single gesture phases as well
as in successions of gesture phases.
Accordingly, our attempt in this paper is to describe gesture phases as part
of a system of movement phases, in which each gesture phase exhibits par-
ticular features that differentiate it from the other gesture phases. We thus ex-
amine the different gesture phases in isolation as well as in reference to each
other. In doing so, we aim at a set of distinctive features based on articulatory
characteristics apparent in the execution of gesture phases, to identify and
characterize the particular phases and to differentiate them from each other.
The focus of this paper is solely on the articulatory characteristics of gesture
phases, which they exhibit independent of speech. With the help of the distinc-
tive features, the different gesture phases will be explored on two levels of
description: one that is context-independent and one that is context-sensitive.
Gesture phases are often embedded within a series of further gesture phases.
Ideally, this series constitutes a gesture unit (Kendon 1980, 2004, see Figure
1), i.e., a succession from a rest position to a rest position containing only one
stroke. This sequence of gesture phases is understood as the context of a ges-
ture phase. In a context-independent description, each gesture phase is de-
picted on its own by means of the distinctive features. Gesture phases are
i solated from their adjacent gesture phases. As each gesture phase shows a
particular realization of the set of distinctive features, each phase can be con-
ceived as a unit of description. In a context-sensitive description, the inuence
adjacent phases may have on a phase is taken into account. For this level of
description, the format of writing phonological rules (Chomsky and Halle 1968)
is adapted to provide for the systematic variations in the set of distinctive fea-
tures characteristic for a type of gesture phase. The context-independent char-
acterization is necessary to perform the context-sensitive characterization as,
rst of all, the units of description need to be determined in order to write rules
that make use of these units.
In a nutshell, it will be shown that 1) gesture phases are describable based on
a limited set of articulatory features, and it is suggested that 2) changes in the
execution of gesture phases due to their sequential embedding can be a ccounted
for based on the set of these distinctive features. The paper therefore provides
a rst step for objectifying the nature of gesture phases as it explicates their
characteristics.
With the following proposal, we wish to reconsider the categories of the
various gesture phases from present-day gesture studies and arrive at an under-
standing and description of gesture phases that take into account their p articular
characteristics and structures that are independent of speech.
The rst section of the paper provides a short description of how gesture
phases have been determined and characterized, by tracing major descriptions
Bereitgestellt von | TU Chemnitz
Angemeldet
Heruntergeladen am | 26.03.15 10:39
58 J. Bressem and S. H. Ladewig
from its beginning stage to its most recent developments. In the next section,
we comment on the study we conducted in order to show what articulatory
characteristics are used for the depiction of the phases. We will then present a
set of distinctive features of gesture phases and characterize the phases based
on them. This context-independent description of gesture phases will be com-
plemented by a context-sensitive description in the third section of the paper.
Here, we exemplify two cases in which the sequential embedding of the phases
in specic linear successions seems to result in a replacement of features in
the particular phases. The last section will discuss advantages of the proce-
dure presented, point out problematic aspects and give an outlook on possible
extensions.
1. Descriptionandclassicationofgesturephasesingesturestudies
The rst scholar of gesture to make the observation that bodily behavior con-
sists of different movement phases or states was Kendon (1972a) in his paper
“Some relationship between body motion and speech.” In this pioneering
a rticle, Kendon examined various types of bodily movements, and showed that
“the pattern of movement that co-occurs with the speech has a hierarchic orga-
nization which appears to match that of the speech units” (Kendon 1972a:
190). This aspect was further developed in his 1980 article, which particularly
addressed the close relationship of gesture and speech as “the same process
of utterance” (Kendon 1980: 208). While the focus in 1972 was on different
levels of body movements, such as shifting the trunk, legs and head, and hands
and arms, the 1980 paper focused solely on the description of “gesticulation,”
i.e., movements of the hands and arms, and aimed at a characterization of the
structure of gesticulation. The paper addresses the phrasal structure of gestures
alone, and provides a classication of the various movement phases that can be
observed in gesturing and accounts for the organization of gestural movements
with respect to building higher-level units.
Regarding the phrasal structure of gestures, Kendon identies six different
phases observable in the execution of gestural movements, i.e., rest position,
preparation, stroke, hold, retraction/recovery, and partial recovery (see Table 2
for more detail). In addition to the gesture phases, Kendon introduces the con-
cepts of the gesticular phrase and the gesticular unit (Kendon 1980: 212).
In his papers, Kendon shows that gesticulation has a structure of its own,
meaning that separate phases of its execution are observable and describable.
Moreover, he provides the terms needed for a closer description and analysis
of gestural movement patterns. His work marks the beginning of the coding of
gesture phases in gesture research, and sets the stage for further accounts of
Bereitgestellt von | TU Chemnitz
Angemeldet
Heruntergeladen am | 26.03.15 10:39
Rethinking gesture phases: Articulatory features of gestural movement? 59
gestural movement patterns. Following distinctions refer implicitly or explic-
itly to Kendon’s differentiations.
Later contributions on gesture phases focus in particular on specic types of
gesture phases. In this manner, Kita, van Gijn, and van der Hulst (1998)3 clas-
sify holds further and point out that a stroke can a) be preceded by a hold, a so
called pre-stroke hold or b) be followed by a hold, a so called post-stroke hold.
This differentiation is grounded in the functions of these two types of holds
with respect to speech. Whereas a pre-stroke hold “is a period in which the
gesture waits for speech to establish cohesion, so that the stroke co-occurs with
the co-expressive portion of speech” (Kita, van Gijn and van der Hulst 1998:
26), a post-stroke hold is “a way to temporally extend a single movement
stroke so that the stroke and post-stroke-hold together will synchronize with
the co-expressive portion of speech” (Kita, van Gijn and van der Hulst 1998:
26). Furthermore, Kita, van Gijn & van der Hulst distinguish between “inde-
pendent holds,” i.e., holds that can stand by themselves and be a “gestural ex-
pression” on their own and “dependent holds,” which ank a stroke and are
“parasitic to the stroke” because “they arise from the semiotic coordination or
modication of the expression in the stroke” (Kita, van Gijn and van der Hulst
1998: 28). This functional specication of holds is currently almost omnipres-
ent in analyses focusing on coverbal gestures (see for example Queck et al.
2002; Gullberg and Holmquist 2006; Kendon 2004; Kettebekov and Sharma
2001; McNeill 2005, in press; Parrill 2001; Sowa 2006) and has been a major
contribution to the discussion of gesture phases (see also Duncan n.d. for fur-
ther investigations on the hold).
A new turn in the identication of gesture phases is initiated with the seg-
mentation method of the “frame-by-frame marking procedure” introduced by
Seyfeddinipur (2006: 105). Former accounts of gesture phases have usually
focused on the denition of gestures phases and their function in relation to
speech (cf. Kita 1990; Duncan n.d.). Questions on how to segment sequences
of gestural movements or questions on how to identify the particular phases
were usually put aside. Although Kita, van Gijn, and van der Hulst (1998) ad-
dressed the segmentation of movement phases, an actual procedure of how to
determine “onsets and offsets of gesture phases” (Seyfeddinipur 2006: 105) has
long been missing.
Seyfeddinipur’s “frame-by-frame marking procedure” aims exactly at these
questions of the identication and coding of gesture phases. It introduces a
methodological approach to the identication of movement phases and estab-
lishes “unambiguous coding criteria for obtaining consistent and frame-
accurate times of gesture phases” (Seyfeddinipur 2006: 105). For this purpose,
Seyfeddinipur takes advantage of an artifact of videos, namely, the sharpness
of a video image in which the execution of movement becomes apparent in
blurry and clear images. Using this artifact, Seyfeddinipur tells apart three
Bereitgestellt von | TU Chemnitz
Angemeldet
Heruntergeladen am | 26.03.15 10:39
60 J. Bressem and S. H. Ladewig
Table 1. Selective overview of existing differentiations of gesture phases
Kendon (1980: 212) Kita, van Gijn and van
der Hulst (1998: 26, 29)
Duncan (n.d.) McNeill (2005: 31) Seyfeddinipur (2006)
rest position “moment when the limb
is in rest”
“part of the body or the
furniture where the
hands can be
supported”
“position where the
self-adapting body
movements take place”
“static position, e.g., on
the lap, arm rests,
arms are folded in
front of the chest”
preparation “a phase in which the
limb moves away
from its rest position
to a position at which
the stroke begins”
“non-stroke phase that
departs form the
resting position”
(includes location
preparation, hand
internal preparation,
liberating movement)
“the limb moves away
from the rest position
into the gesture space
where it can begin the
stroke”
interrupted
preparation/
stroke
“a dynamic phases
abruptly ended and
the abruptness was
not part of the
depiction”
pre-stroke
hold
“a period in which the
gesture waits for
speech to establish
cohesion so that the
stroke co-occurs with
the co-expressive
portion of speech”
“a temporary cessation of
movement before the
stroke”
Bereitgestellt von | TU Chemnitz
Angemeldet
Heruntergeladen am | 26.03.15 10:39
Rethinking gesture phases: Articulatory features of gestural movement? 61
stroke “excursionary
movement in which
the limb shows a
distinct peaking of
effort”
“phase in which more
force is excerted than
in neighboring phases”
“interval of apparent
gestural effort”
“gesture phase with
meaning; it is also the
phase with effort”
“phase which appeared
to display the
meaning of a gesture”
post-stroke
hold
“a way to temporally
extend a single
movement stroke so
that the stroke and post
stroke hold together
will synchronize with
the co-expressive
portion of speech”
“hand freezes in midair
before starting a
retraction, thereby
maintaining the stroke’s
nal position and
posture”
stroke hold “such movements are
strokes in the sense of
meaning and effort but
occur with motionless
hands”
hold “the hand is held still in
the position it
reached at the end of
the stroke”
“phase in which the hand
is held still” (includes
pre- and post-stroke
hold)
“full hold,” “virtual
hold”
“when the hand/s were
held in a static
position other than
the rest position”
full hold “no detectable
movement
virtual hold “some movement
but maintenance
of hand shape
and/or general
location in gesture
space”
Bereitgestellt von | TU Chemnitz
Angemeldet
Heruntergeladen am | 26.03.15 10:39
62 J. Bressem and S. H. Ladewig
Table 1. (Continued )
Kendon (1980: 212) Kita, van Gijn and van
der Hulst (1998: 26, 29)
Duncan (n.d.) McNeill (2005: 31) Seyfeddinipur (2006)
partial
retraction
“phase in which the
hand does not return
all the way to the
position it was in”
“interrupted retraction” “the hands return to rest” “hands move towards a
potential rest position
(e.g., the lap), but
came to a hold before
the rest position was
reached, thus resting
in an intermediate
position”
retraction “phase in which the
limb is either moved
back to its rest
position or is
readied for another
stroke”
“a non-stroke phase that
arrives at rest position”
Bereitgestellt von | TU Chemnitz
Angemeldet
Heruntergeladen am | 26.03.15 10:39
Rethinking gesture phases: Articulatory features of gestural movement? 63
types of transitions in the execution of gestural movement sequences: 1) “tran-
sition from a dynamic to a static phase,” 2) “transition from a static to a dy-
namic phase,” and 3) the “transition from a dynamic to a dynamic phase” (Sey-
feddinipur 2006: 105). These types of transitions are characterized by particular
transition points that mark the move from one type of gestural movement
phase, i.e., dynamic or static, to another. Specic successions of clear and
blurry images thus represent particular sequences of gesture phases. These
types of transitions therefore provide the basis for the assignment of gestural
movement phases to a specic type of gesture phase.
The segmentation procedure developed by Seyfeddinipur (2006) marks a
new milestone in the identication and coding of gestures. It addresses the
question of how to mark boundaries of movement phases and provides a
method, which allows for the differentiation and the marking of gesture phases.
By using characteristics of video images, Seyfeddinipur was able to establish
the basis for a detailed, more exact, and inter-individual segmentation proce-
dure of gestural movement phases. This therefore contributes fundamentally to
the practice of how to code gesture phases.
The preceding section has presented a short overview of the description and
coding of gesture phases and has shown that the categories introduced by Ken-
don in 1980 are still valid today. Although changes in the set up of these cate-
gories have been carried out for years (cf. for instance, Kita, van Gijn, and van
der Hulst 1998), the basic differentiation and characterization of gesture phases
remains the same. Major modications have predominantly been made with
respect to technical aspects involved in the coding of gesture phases (cf. Sey-
feddinipur 2006).
Existing work on the approach to gesture phases has so far mainly been fo-
cused on the proof that a) gestures dispose of a structure of their own, i.e., that
gestural movements are characterized by the progression of specic phases, b)
gesture phases are hierarchically organized and c) they coordinate with units at
the speech level.
All in all, however, gestures are for the most part considered and dened by
their interconnection with other gesture phases, whether preceding or follow-
ing. Likewise, Kendon, for instance, denes the preparation as the “phase in
which the limb moves away from its rest position to a position at which the
stroke begins” (1980: 212). Furthermore, researchers only selectively focus on
the phases themselves and only seldom pursue questions on their nature and
possible characteristics. In a lot of cases, gesture phases are dened respec-
tively in relation to speech:
The approach to gesture phraseology and phaseology is in its essencemeaning-driven.
Locating the beginnings and ends of gesture phrases, or locating the gesture stroke
among a movement’s several phases of execution is a matter of how the phases
Bereitgestellt von | TU Chemnitz
Angemeldet
Heruntergeladen am | 26.03.15 10:39
64 J. Bressem and S. H. Ladewig
(m ovement or hold phases) coordinate, in terms of meaning, with units of the co-
occurring speech, and/or with larger-scale discourse meanings currently in play. (Dun-
can n.d.: 7)
What we would like to contribute to existing accounts on gesture phases with
this paper is a form-based perspective that detaches gestures phases from their
relation to speech, their interconnection to each other and concentrates solely
on possible articulatory features of gesture phases. We therefore ask the fol-
lowing questions: Do the phases show particular articulatory characteristics?
Can one perceive articulatory characteristics in the execution of the phases that
differentiate them from each other? Is it possible to identify and characterize
gesture phases on the basis of such characteristics?
We therefore suggest a focus on articulatory characteristics that are visible
in the execution of gesture phases and aim at their characterization on the basis
of these characteristics. By taking this perspective, the approach tries to ac-
count for a characterization of gesture phases independent of the verbal modal-
ity, as well as a context-independent denition of gesture phases, i.e., without
reference to adjacent phases. The paper tries to separate these varying perspec-
tives by pursuing a form-based approach and by investigating the features and
characteristics of gestures phases from a foremost context-independent per-
spective. More importantly, however, it turns the reader’s attention to the na-
ture of the phases themselves.
2. Describinggesturephases
In this paragraph, we introduce a context-independent description and character-
ization of gesture phases. In particular, the approach aims at two aspects: a) to
present articulatory characteristics apparent in the execution of gestures phases
and b) to characterize and dene the gesture phases based on these characteristics.
By pursuing a context-independent perspective on the various gesture
phases, we examine the phases detached from their position within the move-
ment sequence, i.e., isolate each gesture phase from their adjacency to preced-
ing or following gesture phases. This context-independent description aims at
guring out articulatory characteristics visible in the execution of the phases
that can be used to dene each type of gesture phase, and distinguish the dif-
ferent types of phases from each other.
2.1. Study
In order to determine the specic characteristics of the various gesture phases,
a comparative analysis has been undertaken. For this, we examined ten occur-
Bereitgestellt von | TU Chemnitz
Angemeldet
Heruntergeladen am | 26.03.15 10:39
Rethinking gesture phases: Articulatory features of gestural movement? 65
rences of each gesture phase of ten different speakers. As our focus was on a
form-based characterization of the single gesture phases, as units of descrip-
tion, we excluded any relation to adjacent gesture phases as well as to speech.
Accordingly we arrived at the following gesture phases that were perceivable
based solely on their articulatory characteristics: rest position, preparation,
stroke, hold, and retraction.
The data we used for the analysis are taken from video-recorded conversa-
tions (dyadic, triadic, as well as smaller group constellations) of a parlor game,
and conversations, which were not conned to a certain topic (see also Ladewig
2007, 2010, forthcoming). Furthermore, we used recordings of the German TV
show Genial Daneben in which ve comedians try to answer peculiar ques-
tions sent in by viewers. From these recordings, we chose an equal share of
women and men whose gestures we examined with respect to the gesture
phases. The gestures that would build the basis for the analysis were chosen
randomly.
For the rst step, we segmented the gestures based on the “frame-by-frame
marking procedure” (Seyfeddinipur 2006, see also Section 2) into single move-
ment phases. Using this procedure, we dened different onsets and offsets of
movement phases within the ow of gestural movements (cf. Seyfeddinipur
2006: 105). Following Seyfeddinipur’s segmentation procedure, we p artitioned
the gestural ow by marking changes in the execution of movement. Using the
annotation software ELAN4 we were able to identify static and dynamic phases
from the ow of gestural movement. We then assigned the segments, which we
obtained based on the “frame-by-frame marking procedure,” to the different
gesture phase categories.
In order to determine the distinguishing features of the gesture phases, we
then analyzed the ten occurrences of each gesture phase in three separate steps.
First, we looked at the instances of each gesture phase and watched them
repeatedly in order to get a feeling of what kinds of observable articulatory
attributes the gesture phases exhibit.
The re-viewings permit a search for the natural usages and regularities within that pro-
cess [the ow of human interaction]. The discovery of patterns leads, in time, to the
clarication of unit identities within that process. The system itself, instead of an infor-
mant, provides the contrast which permits decisions. Analysis proceeds over time and
is cumulative, with each bit of new evidence supporting previous ndings and provid-
ing clues about further patterns. (Condon and Ogston 1967: 223)
Using this discovery procedure, we noted down striking articulatory character-
istics in the appearance of the different phases, which we observed while
watching them. In this rst step of the analysis, we did not exclude any articu-
latory characteristic in the phases’ appearance but wrote them down without
making a choice.
Bereitgestellt von | TU Chemnitz
Angemeldet
Heruntergeladen am | 26.03.15 10:39
66 J. Bressem and S. H. Ladewig
While watching the instances repeatedly we noticed for example that, in
most cases, the hands would move upwards in a straight movement pattern
during a preparation and usually straight downwards in a retraction. Only sel-
dom did we observe a different direction (e.g., preparation downwards and
retraction upwards). In these cases, the hands of the speakers were above head
level. Accordingly, regarding possible articulatory features of a preparation or
retraction, in this rst step of the analysis, we noted straight upwards and
downwards movements.
In a second step, we compared the articulatory characteristics of the gesture
phases with each other and picked out similarities and differences. The goal
was to nd out which characteristics occurred in all of the different phases and
which ones were only observable in particular phases, and to sort the charac-
teristics into groups of possible articulatory characteristics. Coming back to the
example of the preparation and retraction given above, the particularity of the
straight movement upwards or downwards was sorted as a feature occurring
preferably in selective phases (see Section 2.2.2. on additional features).
In a third step, we then classied the observed qualities into a) distinctive
features and b) additional features (see below).
2.2. Articulatory characteristics of gesture phases: A context-independent
description
Based on the study presented above, we arrived at two types of articulatory
features characterizing gesture phases. In particular, we dened two sets of
features: a) distinctive features and b) additional features. The set of distinctive
features comprises the categories “movement” and “tension.” The set of addi-
tional features consists of the categories “possible types of movement” and
“ow of movement.”
Although both the distinctive and additional features are based on articula-
tory characteristics, which are observable in the execution of the phases, the
two sets of features show different distributions. The set of distinctive features
compromises attributes visible in all gesture phases. Accordingly, each gesture
phase realizes features from this set. However, each gesture phase shows a
particular implementation of these features.
The distinctive features make up a paradigmatic set of properties. The fea-
tures included in one category, as, for instance, in the category “movement,”
are mutually exclusive of one another and stand in opposition to each other. If
a gesture phase carries one particular distinctive feature, it cannot carry an-
other feature from the same category. This is because the existence of one of
the features results in the exclusion of the other ones from the same category.
Hence, the distinctive features identify the phases as specic and differentiate
Bereitgestellt von | TU Chemnitz
Angemeldet
Heruntergeladen am | 26.03.15 10:39
Rethinking gesture phases: Articulatory features of gestural movement? 67
the phases from each other. Each gesture phase shows a distribution of the
distinctive features that is particular for only this type of phase.
Apart from the distinctive features, we were able to identify further traits of
gesture phases. Contrary to the distinctive features presented above, these ad-
ditional features cannot be observed in all gesture phases. They only apply to
phases that exhibit the distinctive feature “presence of movement.” As these
features only apply to specic phases and are not discernable in every instance
of a phase that shows the feature “presence of movement,” they can only be
used to identify gesture phases selectively. They will be used for a further de-
scription in order to enhance the formal account of the gesture phases to be
presented. Both types of articulatory features will be introduced in the follow-
ing section.
2.2.1. Distinctive features of gesture phases. Based on the study presented
above, we were able to identify two categories of distinctive features that char-
acterize the execution of gestural movement sequences. These features are:
– movement
– tension of the hand
Movement. Probably the most prominent characteristic observable in the exe-
cution of gesture phases is the movement of the hands and arms. Gesture
phases can thus be distinguished according to the presence or absence of move-
ment. Accordingly, the category movement consists of the features:
a) presence of movement, [+movement], and
b) absence of movement, [−movement].
Tension. The category tension captures the tenseness of the hands during the
execution of a gesture phase and aims at changes between phases of relaxation
and exertion. During the execution of the different gesture phases, the hands
are characterized by a variation in the occurrence and strength of tension. It is
possible to observe phases in which the tension increases or decreases. At other
times, the tension remains stable throughout the execution of a phase.5
Changes in tension of the hand can be observed from three kinds of m uscular
activities in different parts of the body: the volar and dorsal exion of the n-
gers, the hand as well as the rotation of the arm. The volar exion refers to the
bending of the ngers towards the palm, the dorsal exion to their extension.
The hand conguration in these states of tension differs from that assumed in
the default condition, which is characterized by residual muscle tension (tonus,
Figure 2). In this ( basic) condition, the muscles contract continuously and help
maintain the body’s posture. Thus, minimal tension is produced at all times.
Bereitgestellt von | TU Chemnitz
Angemeldet
Heruntergeladen am | 26.03.15 10:39
68 J. Bressem and S. H. Ladewig
As tension can, among others, be observed in the exion of the single n-
gers, it is self-evident that different qualities in tension are reected in the
conguration of the hand. More specically, in the different stages of forma-
tion and deformation as well as in the stage of being fully formed.
The category “tension” shows two features:
a) presence of tension, [+tension], and
b) absence of tension, [−tension].
The feature [+tension] applies when the ngers either stretch (dorsal exion)
or bend towards the palm (volar exion). The feature [−tension] applies when
the hands are relaxed in a default condition (see above and Figure 2).
Phases showing the feature [+tension] can be sub-classied and marked by
the features [+constant] and [−constant]. These features reect transitions in
tenseness, which become visible in transitions of the hand’s conguration, i.e.,
formation, deformation or a stable hand conguration.
a) Constant [+constant]: In movement phases characterized by the feature
[+constant] the beginning and the end of a phase does not show a difference in
Figure 2. Tension visible in the hand’s conguration
Bereitgestellt von | TU Chemnitz
Angemeldet
Heruntergeladen am | 26.03.15 10:39
Rethinking gesture phases: Articulatory features of gestural movement? 69
tenseness. In these phases a specic hand conguration, which differs from the
ones that can be taken in the default condition is maintained.
b) Not constant [−constant]: In movement phases characterized by the feature
[−constant], the tenseness of the hand either increases or decreases. In terms of
visible characteristics, the starting point of such a phase differs from the end-
point, in that the former marks the beginning of the exion of the ngers
whereas the latter marks the reaching of the full form. Accordingly, an increase
is reected in the formation of the hand’s conguration. In phases identied by
decreasing tension, the opposite is the case. The beginning of the phase is char-
acterized by the hand’s tenseness, whereas the endpoint exhibits a decline of
tenseness. The decrease in hand tension corresponds with a deformation of the
hand’s conguration approximating the default condition (see also Harling and
Edwards 1997; Martell 2005).
Thus phases characterized by the feature [−constant] can be further sub-
classied. Thus increasing tenseness, visible in the formation of a congura-
tion, and decreasing tenseness, visible in the deformation of a conguration,
are both captured by the feature pair [+-increase]. Accordingly the former is
marked by [+increase], the latter by [−increase] (see Figure 3).
As has been pointed out, the features of the category “tension” become vis-
ible in the conguration of the hand, which differs from congurations that
could be taken in the default condition (see Figure 2). Two different kinds of
m uscular activities contribute to this difference — the volar and dorsal exion
of the ngers. Accordingly, the conguration of the ngers is consulted in
order to account for changes in tension.
However, in some cases it is possible to apply the feature [+tension] for
which no exion of the ngers can be observed. In these cases, tension is
Figure 3. Category “tension”
Bereitgestellt von | TU Chemnitz
Angemeldet
Heruntergeladen am | 26.03.15 10:39
70 J. Bressem and S. H. Ladewig
r eected in the movement of the hand, i.e., the movement is anchored at the
wrist, and/or arm:
a) The hand can be bended towards the pulse (volar exion).
b) The hand can be raised towards its back (dorsal exion).
c) The arm and thereby the hand can be rotated. Supination refers to the rota-
tion upward, pronation to the rotation downwards.6
Although these cases can be observed, most of the time the exion of the n-
gers goes along with a exion and/or rotation of the hand.
With the help of these distinctive features the different movement phases
identied in the “frame-by-frame marking procedure” (Seyfeddinipur 2006)
can be dened as gesture phases by relying on articulatory characteristics
alone. The additional features that characterize the different gesture phases
more closely can enhance this formal account. These features will be elabo-
rated in the following section.
2.2.2. Additional features. Dynamic gestures phases, i.e., phases character-
ized by the feature [+movement], show additional articulatory qualities that
can be used to buttress the identication of dynamic gesture phases, and to
further differentiate this class of gesture phases from each other. Accordingly,
with respect to additional features of dynamic gestures phases, i.e., gesture
phases carrying the feature [+movement], we distinguish two additional cate-
gories. Here, we introduce the category “type of movement” with its features
[+-restricted] and the category “ow of movement” with the feature [+-vari-
able]. Features from these two sets are used to buttress the identication of the
gesture phases and provide additional articulatory attributes of dynamic ges-
ture phases. The features subsumed in these two sets however partially differ
from the distinctive features introduced in the section above. The distinctive
features apply to all gesture phases and thus generally distinguish segments
from each other because a segment cannot be marked as a [+movement] and a
[−movement] at the same time. The additional features are only applicable to
dynamic gesture phases. Accordingly, these features are not common to ges-
ture phases in general, but rather to a subset of phases.
However, regarding the differentiation of the dynamic gesture phases, the
additional features carry comparable weight as the distinctive features. For the
dynamic phases themselves, they have a distinctive function as they clearly
show a particular distribution across the dynamic phases and differentiate the
phases from each other. The same is true for the distinctive features. The fea-
tures from the two categories of the additional features are exclusive of each
other, meaning that for example one phase cannot carry the features [+vari-
able] and [−variable].
Bereitgestellt von | TU Chemnitz
Angemeldet
Heruntergeladen am | 26.03.15 10:39
Rethinking gesture phases: Articulatory features of gestural movement? 71
2.2.2.1. Possible types of movement. In the data, we found that the gesture
phases vary according to the types of movement that can apply. These differ-
ences are captured by the category “types of movement” (cf. Bressem 2008).
With respect to the phases characterized by the execution of movement, two
features are distinguishable:
a) Restricted [+restricted]: Gesture phases that are restricted in their range of
types of movement show either “straight” or “curved” movements. When the
movement is anchored solely at the wrist, bending and raising of the hand as
well as rotation can be observed.
b) Not restricted [−restricted]: Gesture phases that are not restrained in their
range of movement can show the following: 1) “straight,” 2) “curved,” 3)
“c ircle,” 4) “spiral,” 5) “zigzag,” and 6) “s-line.” Furthermore, movements of
the wrist, such as “bending,” “raising,” and “rotation” may be performed.
2.2.2.2. Flow of movement. Gesture phases characterized by the execu-
tion of movement can show a variation in the quality of movement. The mo-
tion may be executed with varying force and/or a change of velocity. The
c ategory “ow of movement” captures these differences.8 Two features are
distinguished:
a) Variable [+variable]: The ow of movement can be described as variable if
it shows some degree of variation within one movement phase: It may be ac-
centuated, accelerated or decelerated (Bressem 2008).
b) Not variable [−variable]: If the ow of movement does not show any varia-
tion the feature [−variable] is applied.
The following table gives an overview of the distinctive and additional features
described in this section.
Table 2. Overview of distinctive and additional features
distinctive features additional features
possible types of
movement
ow of
movement
movement +movement +restricted
−restricted
+variable
−variable
−movement
tension +tension +constant
−constant +increase
−increase
−tension
Bereitgestellt von | TU Chemnitz
Angemeldet
Heruntergeladen am | 26.03.15 10:39
72 J. Bressem and S. H. Ladewig
In this section, articulatory characteristics were described that were arrived
at in a context-independent description. They can be used to identify and dene
movement phases in terms of gesture phases. A set of distinctive and a set of
additional features have been determined based on the evaluated data. The set
of distinctive features consisting of the categories “movement” and “tension”
can be observed in all gesture phases. They are therefore used for identifying
the ve gesture phases rest position, preparation, stroke, hold, and retraction.
The set of additional features can only be applied to such gesture phases that
are characterized by the execution of movement ([+movement]). The features
distinguished in the categories “types of movement” and “ow of movement”
can be used to buttress the identication of gesture phases by the distinctive
features.
Below the ve gesture phases identied in the data are dened by using the
sets of distinctive and additional features. Again the description of these phases
relies only on articulatory characteristics and not on functional properties.
2.3. Description of gesture phases
As already mentioned in the section above, we are distinguishing ve basic
gesture phases, which can be identied and described based on observable ar-
ticulatory characteristics. These are: 1) rest position, 2) preparation, 3) stroke,
4) hold, and 5) retraction.
Based on the category “movement,” these phases can be divided into two
groups: a) phases that involve the execution of movement, i.e., dynamic ges-
ture phases and b) phases that do not show movement, i.e., static gesture
phases. The following section will rst describe dynamic gesture phases and
then proceed to describing static gesture phases.
2.3.1. Preparation. The gesture phase preparation is characterized by the
execution of movement. During the performance of a preparation, the tense-
ness of the hand increases, which is reected in the formation of the hand’s
conguration. This means that the hand assumes a conguration that differs
from the one that can be assumed in a default condition. Accordingly, this ges-
ture phase is marked by the features [+movement], [−constant], and [+increas-
ing]. The variation of the possible types of movement is restricted, meaning
that only “straight” and “curved” (Bressem 2008) movements come into ques-
tion, whereby the former is performed most often. If the movement is anchored
merely at the wrist of the hand, the variation of movement is restricted in these
cases as well: The hand can only be raised (dorsal exion) and/or rotated (su-
pination). Accordingly, the preparation carries the feature [+restricted]. The
Bereitgestellt von | TU Chemnitz
Angemeldet
Heruntergeladen am | 26.03.15 10:39
Rethinking gesture phases: Articulatory features of gestural movement? 73
ow of movement does not show any variation, i.e., the feature [−variable] is
applied.
2.3.2. Retraction. The gesture phase retraction is also characterized by the
execution of movement. During the performance of a retraction the tenseness
of the hand decreases and the hand’s conguration is modied, insofar as its
formation is resolved and approximates a default condition. Accordingly, this
gesture phase is marked by the features [+movement], [−constant], and [−in-
creasing]. The variation of the possible types of movement is restricted: only
“straight” and “curved” (Bressem 2008) types of movement apply whereby the
former is performed most often. If the movement is anchored solely at the
wrist, the hand can only be rotated ( pronation) or bent to pulse (volar exion).
The retraction thus carries the feature [+restricted]. With respect to the ow of
movement, the feature [−variable] is applied.
In order to facilitate the distinction of the gesture phases preparation and
retraction, the distinctive as well as additional features are contrasted in
Table 3.
As can be seen in the table, both phases differ from each other only with
respect to the category “tension.” Whereas the tenseness of the hand increases
during the performance of the preparation, it decreases in the implementation
of a retraction, which is reected in different formations of the hand.
2.3.3. Stroke. A gesture’s stroke is characterized by the execution of move-
ment. The tenseness of the hand remains stable. Accordingly, in most cases the
conguration of the hand does not change. In some cases, the shape of the hand
may be transformed during the performance of a stroke, e.g., a nger is raised
during the execution of a circle. But as these changes occur within one move-
ment phase and since the tension may undergo changes only in parts of the
hand, the tension of the whole hand cannot be considered as resolved. Accord-
ingly, even in these instances the tension can be regarded as constant.
The stroke is marked by the features [+movement] and [+constant]. Unlike
the gesture phases characterized above, the types of movements are not
Table 3. Contrast of preparation and retraction
Feature Movement possible types
of movement
ow of
movement
Tension
Phase
Preparation [+movement] [+restricted] [−variable] [−constant] [+increase]
Retraction [+movement] [+restricted] [−variable] [−constant] [−increase]
Bereitgestellt von | TU Chemnitz
Angemeldet
Heruntergeladen am | 26.03.15 10:39
74 J. Bressem and S. H. Ladewig
r estricted at all. The gesture phase stroke is the only phase characterized by
movement, which is not restrained in the implementation of motion patterns
(see also Seyfeddinipur 2006: 106). In particular, the following six basic types
of movement may be realized: 1) “straight,” 2) “curved,” 3) “circle,” 4)
“s piral,” 5) “zigzag,” and 6) “s-line” (Bressem 2008). Furthermore, move-
ments anchored at the wrist, such as “bending,” “raising,” and “rotation”
(Bressem 2008) may be performed. In addition, the movements executed dur-
ing a stroke may be accentuated and/or performed with changing velocity.
Thus, the stroke can be further marked as [−restricted] and in some cases as
[+variable]. These features do not apply to any of the further gesture phases.7
As Table 4 shows, the stroke is the gesture phase among those characterized
by movement in which the hand is tensed throughout the phase. Furthermore,
this phase may exhibit an invariable ow of movement.
2.3.4. Hold. The hold is one of the gesture phases that is marked by the lack
of movement and thus clearly differs from the phases characterized above. Ac-
cordingly, the feature [−movement] is applied. The hand is tensed throughout
the execution of a hold, meaning that the hand’s conguration is maintained.
The conguration thus differs from the ones that can be assumed in the default
condition, thus showing the features [+tension] and [+constant]. Sometimes a
hold can show some movement, i.e., a slight drifting observable in its e xecution
(see Duncan n.d.: 4; for a further differentiation of holds, see Tag in p reparation).
Stroke and hold share one distinctive feature that is not realized in one of the
remaining phases and which is essential for their identication — [+tension]
that is further sub-classied by the feature [+constant].
Table 4. Contrast of preparation, stroke, and retraction
Feature Movement possible types
of movement
ow of
movement
Tension
Phase
Preparation [+movement] [+restricted] [−variable] [−constant] [+increase]
Stroke [+movement] [−restricted] [+variable] [+constant]
Retraction [+movement] [+restricted] [−variable] [−constant] [−increase]
Table 5. Contrast of stroke and hold
Feature Movement possible types of movement ow of movement Tension
Phase
Stroke [+movement] [−restricted] [+variable] [+constant]
Hold [−movement] [+constant]
Bereitgestellt von | TU Chemnitz
Angemeldet
Heruntergeladen am | 26.03.15 10:39
Rethinking gesture phases: Articulatory features of gestural movement? 75
2.3.5. Rest position. A rest position is also characterized by an absence of
movement. The hands usually do not show any tension. This gesture phase is
marked by the features [−movement, −tension]. If the hands are moved, then
they either touch a part of the speaker’s own body or an external object, such
as a table or cup. These movements have been described as self-adaptors,
o bject-adaptors, and dgeting (see, e.g., Ekman and Friesen 1969, 1974, 1977;
Freedman 1972; Müller 1994, 1998). Any movements that belong to the cate-
gory of “self-touch” or object manipulation were not included into the descrip-
tion and are considered as non-gestural. Self-touching movements are move-
ments such as scratching, hair stroking, nose rubbing, etc. Object manipulation
includes holding a cup, dgeting with a necklace or a pen, etc. In cases in
which the body or object functions as reference object, we speak of “body-
focused movements” (Freedman 1977) or “object-focused movements”8
(Freedman 1977). We are aware of the fact that further research is needed to
provide articulatory characteristics in order to distinguish such movements
from body-focused movements or object-focused movements. However, as
has been shown recently, people are able to distinguish dgeting from gestural
movements: “[S]igns, as cases of gesture, can be discriminated from other
movements, in our study dgeting, by their appearance.” (Arendsen, Doorn,
and Ridder 2007: 330; see also Kendon 2004).
The section above has presented the characterization of the ve gesture
phases rest position, preparation, stroke, hold, and retraction on the basis of
distinctive as well as additional features, which are based on articulatory fea-
tures observable in the execution of these phases. We have provided denitions
for each phase based on these features. We discussed the distribution of the
features in each gesture phase as well as across phases, presented characteristic
sets of features for each type of gesture phase, and set up a feature matrix.
This characterization however constitutes an ideal depiction of the various
gesture phases, meaning that we have described the phases with respect to the
features they usually exhibit. These characteristics can undergo changes in par-
ticular linear sequences of gesture phases. The following section will thus
present two types of sequential embeddings of the phases in specic linear suc-
cessions, which result in changes of the articulatory characteristics of p articular
phases.
3. Articulatorycharacteristicsofgesturephasesinlinearsuccessions
—acontext-sensitivedescription
When watching people gesture, one observes that gestures and therefore ges-
ture phases follow each other immediately. We can speak of an ideal succes-
sion of gesture phases if the speaker’s hands progress from a rest position to a
Bereitgestellt von | TU Chemnitz
Angemeldet
Heruntergeladen am | 26.03.15 10:39
76 J. Bressem and S. H. Ladewig
preparation, then execute a stroke and afterwards progress via a retraction to a
rest position again (see Figure 4).
However, this ideal succession of gesture phases constitutes just one possi-
bility of how gesture phases can follow each other in order to constitute larger
gestural units, such as gesture phrases or gesture units. In many cases, for
example, rest positions are missing, meaning that the hand after executing a
stroke or a retraction does not return to its rest position, but rather sets off in
order to perform a new preparation or even a stroke. Also, preparations are
quite frequently missing, so that strokes follow each other immediately with-
out exhibiting preparational phases in between.
In a lot of these instances in which the execution of gestures “deviate” from
the ideal succession of gesture phases, changes in the articulatory characteris-
tics of the phases can be observed. Due to their sequential embedding in a
specic linear progression, the phases’ characteristics might undergo changes.
In a preparation, for example, usually characterized by the features [+move-
ment, +tension, +increase], it is possible to observe instances in which the
feature [+increase] is replaced. In other instances, differences in types of rest
positions, such as hands resting in the lap or on the body of the speaker, result
in changes of the foregoing retraction phase and the replacing of the feature
[+decrease]. With gesture phases, in particular linear successions, it is there-
fore possible to observe changes in their articulatory features.
These changes resemble coarticulation phenomena in spoken language,
namely, instances in which neighboring sounds inuence each other in their
a rticulation (see for example Menzerath and Lacerda 1933; Trubetzkoy 1958).
Considering the similarities in the articulation process of speech and gesture, the
existence of coarticulation phenomena in coverbal gestures is not surprising (cf.
Fricke 2008). The same is true for speech. The articulation of coverbal gestures
is a continuing process in which the separation of the individual gesture phases
is not unproblematic, as the different phases seem to merge into each other.
In producing speech, the vocal organs “perform gliding uent fading move-
ments in which the positions for particular sounds are not realized in a down-
right xed manner, but are rather aimed a relative manner” (Ternes 1999: 37,
Figure 4. Ideal succession of gesture phases
Bereitgestellt von | TU Chemnitz
Angemeldet
Heruntergeladen am | 26.03.15 10:39
Rethinking gesture phases: Articulatory features of gestural movement? 77
our translation). This gliding into each other of the articulation movements
brings about the inuencing of sounds on one another. In the case of the Ger-
man words Kies (gravel) and Kuss (kiss), for example, the pronunciation of the
sound [k] differs. In the case of Kies, the articulation of the [k] is fronted be-
cause of the anticipation of the vowel [i:]. In the case of Kuss, however, the [k]
is pronounced more in the back because it precedes a back vowel. However,
for the treatment and especially the identication and coding of gesture phases
such coarticulation phenomena pose a major problem. The difcult identiable
boundaries of the particular phases or segments are thus even catchier to per-
ceive and manage.
By taking two particular cases in which the execution of the phases change,
we would like to show that based on the articulatory characteristics introduced
in this paper a context-sensitive description of phases in linear succession is
possible, which shows that these changes do not show up randomly, but are
rather limited to particular sequential contexts.
In order to describe changes in the features sets of the gesture phases, we
will use the format of phonological rules as examined by Chomsky and Halle
(1968). Phonological rules describe phonological changes in words and make
general statements about the relations between sounds or types of sounds.
They summarize the behavior of sounds in particular phonetic and g rammatical
surroundings. Chomsky and Halle (1968) distinguish between three different
types of phonological rules: rules describing the insertion or deletion of a pho-
nological segment, and rules describing changes in the value of segments. (The
latter will be of interest for the context-sensitive description of gesture phases.)
What is important to point out is that phonological rules do not substitute
whole segments, but only individual features and are thus especially capable of
capturing coarticulation phenomena.
An example for the use of phonological rules in linguistic analyses is the
description of nal devoicing in German. In German, the opposition between
voiceless and voiced obstruents is neutralized in the nal sounds of words,
such as Rad (wheel) are pronounced in the same way as Rat ( board or council),
i.e., as /ra:t/. Written in terms of a phonological rule, nal devoicing in German
can be noted as follows:
/ b d g v z ʒ/ → [ p t k f s ʃ ]/ #.
Figure 5 shows the general scheme for the notation of phonological rules. The
rule of nal devoicing in German reads as follows: The phonemes / b d g v z ʒ/
make up the input of the rule and [ p t k f s ʃ ] the output of the rule. The arrow
denotes “is realized as.” The / stands for “in the context of.” The underline
species the place where the changes takes places and the # stands for the end
of a word. In sum, the rule describes the fact that the phonemes / b d g v z ʒ/ are
realized as [ p t k f s ʃ ] at the end of a word in German.
Bereitgestellt von | TU Chemnitz
Angemeldet
Heruntergeladen am | 26.03.15 10:39
78 J. Bressem and S. H. Ladewig
However, before focusing on the description of the gesture phase examples,
we would like to clarify that the formulae are used solely as a tool to capture
changes in the features of gesture phases, in particular linear environments. We
use the formulae to describe observable phenomena in the execution of gesture
phases, and do not adopt any of the implications of the theory of language as-
sociated with the use of phonological rules in generative grammar (cf. Chom-
sky 1993). Furthermore, we do not imply that the changes in the features are
based on an underlying gesture phase feature structure. We use this notation
as it is widely accepted in phonology and best captures the phenomena under
investigation.
In our comparative study, we observed linear successions of gesture phases
with a diverse range of modications, meaning that the types and number of
segments in which features are replaced differed. In general, we were able to
distinguish linear sequences that cause replacement of features in a) one ges-
ture phase or b) in two or more gesture phases (see Bressem in preparation).
Whereas in the former, only one gesture phase shows replacement of some of
its features, the latter causes replacement in the features of two or more adja-
cent gesture phases. In the following section, we will not discuss the second
type but rather concentrate on two examples in which features of one gesture
phase are replaced due to the sequential context.
3.1. No execution of preparation
As was mentioned in the beginning of this section, in an ideal linear succession
of gesture phases (see Figure 3 above), a stroke is preceded by a preparation.
Speaking in terms of the features presented, a stroke is thus usually preceded
by a segment with the features [+movement, +tension, −constant, +increase,
+restricted, −variable].
However, instances in which the preparation is dropped are quite common
(see Kendon 1980, 2004; Kita, van Gijn, and van der Hulst 1998; McNeill
1992; Seyfeddinipur 2006). In the course of the comparative study as well
as previous research on gesture phases, we particularly noted this type of
Figure 5. General schema for the notation of phonological rules
Bereitgestellt von | TU Chemnitz
Angemeldet
Heruntergeladen am | 26.03.15 10:39
Rethinking gesture phases: Articulatory features of gestural movement? 79
phenomenon for specic occurrences of deictic gestures (see Figure 6 for an
e xample).
What is striking about these instances with respect to distinctive features of
gestures phases as presented in the foregoing section, is the fact that, due to the
omission of the preparation, the stroke undergoes changes in some of its dis-
tinctive features. A stroke is usually characterized by the features [+movement,
+tension, +constant, −restricted, +variable] (see Table 5). A stroke in a linear
succession missing a preparation may change in the category “tension” as well
as in one of the additional form features.
In sequences in which no preparation is performed, as in the case of (spe-
cic) deictic gestures, the stroke must now be characterized in the following
way [+movement, +tension, −constant, +increase, +restricted, +variable].
Contrary to the context-independent characterization, the stroke in successions
Figure 6. Example of no preparation in the succession of gesture phases
Bereitgestellt von | TU Chemnitz
Angemeldet
Heruntergeladen am | 26.03.15 10:39
80 J. Bressem and S. H. Ladewig
without a preparation is now marked [−constant] and is further specied for the
feature [+increase] as well as [+restricted]. Contrary to the usual stroke, a
stroke in these successions shows an increase of tension instead of the tension
being constant. Also, it is restricted in the types of movements that are execut-
able, such that mostly straight and curved movements occurred. All in all, the
stroke thus now carries features that are distinctive for a different gesture
phase, namely, a preparation (see Section 2.3).
The omission of the preparation thus has an impact on the following gesture
phase, namely, the stroke, and leads to the replacement of particular distinctive
features of that movement segment. These types of replacements have so far
been observed only in successions in which the stroke is either preceded by a
rest position or a retraction. So, the context in which the features of a stroke
can be replaced seems to be limited.
However, although the stroke shows no distinctive features characteristic for
a preparation, the stroke is still clearly identiable as such as it is marked for
the feature [+varying], a feature it does not share with any other gesture phase.
So, in the instances investigated, the ow of movement executed in the stroke
phase could vary. In quite a few instances, the end of the movement was ac-
centuated (see Bressem in preparation),9 a characteristic of movement phases
that can only be found in strokes.
In order to describe the replacement of features visible in these particular
instances of strokes, we will apply the format of phonological rules (cf. Hall
2000). By doing so, we are able to account for the replacement of individual
features in the particular contexts that we observed. It allows us to relate the
context-independent description presented in the previous section with the
context-sensitive description, and lets us account for the change of gesture
phase characteristics in linear successions.
The replacement of features in the stroke observed in the particular cases of
missing a preparation can thus be formulated as shown in Figure 7.
What we would like the reader to bear in mind is that this description along
with the rules we presented, is not meant to be universally valid in all instances
of missing preparations. Rather, what we presented is the rst description of
one particular phenomenon visible in the execution of gesture phases of spe-
cic deictic gestures. Whether similar effects can be described for other in-
stances in which preparational phases are omitted remains to be answered
(Bressem in preparation).
3.2. Transition from a retraction to a rest position on the speakers’ body or
to an action
Another interesting example of changes in the distinctive features caused by
the linear succession, in which the gesture phases are em bedded, is the transi-
Bereitgestellt von | TU Chemnitz
Angemeldet
Heruntergeladen am | 26.03.15 10:39
Rethinking gesture phases: Articulatory features of gestural movement? 81
tion from a retraction to a rest position on the speaker’s body or to an action. In
these cases, the speaker does not bring his or her hands into a typical rest posi-
tion (e.g., lap, table, see Figure 8) but rather, after executing a retraction, moves
his hands into a rest position on his or her body or starts to perform actions.
Examples of rest positions on the speaker’s body can be seen in Figure 8. Ac-
tions that can be performed after a retraction are self-stimulating movements or
taking hold of objects (see Section 2.3).
Figure 7. Rule expressing the changes of distinctive features in the gesture phase stroke in
instances of no preparation
Figure 8. Possible rest positions on the speaker’s body10
Bereitgestellt von | TU Chemnitz
Angemeldet
Heruntergeladen am | 26.03.15 10:39
82 J. Bressem and S. H. Ladewig
Before making further descriptions of the cases in mind, we would like to
point out that the types of instances we are referring to do not include cases in
which the speaker indexes to him/ herself. Cases such as a deictic gesture with
a stretched index nger on the chest of the speaker for example do not fall
under the phenomenon presented here (see Section 2.3 for body-focused move-
ments). We only include instances in which the retraction of the hand leads to
a rest position or the performance of an action.
In these instances again, particular distinctive features of the gesture phase
retraction seem to be replaced by other features. Similar to the example in
which the preparation is omitted, thus causing a replacement in features of the
stroke, the retraction also shows the replacement of some of its distinctive
features. Usually a retraction is characterized by a decrease in tension along
with a deformation of the hand’s conguration and has been dened in section
2 on the basis of the distinctive features as follows: [+movement, +tension,
− constant, −increase, +restricted, −variable]. However, in cases in which the
hand transitions into rest positions on the body or into actions after the execu-
tion of a stroke, the feature [−constant] is replaced. The tension no longer de-
creases, but rather stays constant. The constant tension furthermore goes along
either with the maintenance or the change of a hand’s conguration. In these
cases, the retraction can no longer be characterized as [+movement, +tension,
−constant, −increase, +restricted, −variable] but has to be accounted for with
the features [+movement, +tension, +constant, +restricted, −variable]. Accord-
ingly, retractions in these linear successions now carry a distinctive feature that
is an actual characteristic for strokes and holds, i.e., [+constant]. However, due
to the marking of the feature [−variable], the movement segment cannot be
mistaken as either one of the two phases and thus must be identied as a ges-
ture phase usually characterized by changes in tension.
By means of the format of phonological rules (see Figure 9), we are once
again able to capture the changes in the particular distinctive feature of retrac-
tions. Similar to the cases of a missing preparation, this format allows us to
describe the observable changes in the retraction in such transitions, based on
the distinctive features introduced in the foregoing sections.
In this section, we presented two cases in which the sequential embedding
of the phases in specic linear successions may result in replacements of par-
ticular distinctive features. An examination of two instances has shown that the
replacement of distinctive features only affects individual features and seems
to depend on particular contexts.
Both cases presented thus appear to reect the fact that the articulation of
gestures, similarly to the verbal utterance, is a continuing process in which the
different phases glide into each other are mutually affecting their articulation.
As a result, coarticulation phenomena, such as the ones presented in this sec-
tion, can be observed on the level of gesture phases. However, the kinds of
Bereitgestellt von | TU Chemnitz
Angemeldet
Heruntergeladen am | 26.03.15 10:39
Rethinking gesture phases: Articulatory features of gestural movement? 83
coarticulation phenomena that come into play here cannot be answered yet and
go beyond the scope of this paper. For answering questions on the direction
and range of change, such as whether the cases presented here could be con-
ceived of as examples for progressive or regressive assimilation (cf. Trubetz-
koy 1958) needs further investigation. The aim of this paper was to provide a
rst contribution to this question, and to show that based on distinctive features
dened on articulatory characteristics of gesture phases, such phenomena can
be described.
4. Summary
In this paper the denition of gesture phases applied in current gesture studies
has been reconsidered. A context-independent depiction of the phases prepara-
tion, stroke, hold, retraction, and rest position based on their articulatory char-
acteristic alone, leaving functional aspects aside, was proposed.
Based on a comparative analysis of the different gesture phases, we pro-
vided two sets of features, which are based on articulatory characteristics ob-
servable in the execution of these phases, namely, distinctive features and ad-
ditional features. Besides providing denitions for each of the ve gestures
phases based on these features, we discussed the distribution of the features in
each gesture phase as well as across phases. According to this, characteristic
sets of features for each type of gesture phase were presented and a feature
matrix was set up. With respect to the distinctive features, two categories were
proposed — “movement” and “tension.” Features from these categories apply
to all gesture phases but their distribution varies, i.e., each gesture phase shows
a distribution of these features that is particular for only this type of gesture
phase. Phases characterized by the feature marked [+tension] can be sub-
Figure 9. Rule expressing the changes in formal features of a retraction in transitions to rest
positions on the body of the speaker as well as transitions to actions
Bereitgestellt von | TU Chemnitz
Angemeldet
Heruntergeladen am | 26.03.15 10:39
84 J. Bressem and S. H. Ladewig
classied according to transitions in the tenseness of the hand, which become
visible in transitions in the hand’s conguration (feature pairs [+-constant] and
[+-increasing]). Additional features only apply to dynamic gesture phases, i.e.,
phases characterized by the feature [+movement]. The categories that make up
the set of additional features are “types of movement” and “ow of move-
ment.” The category “types of movement” shows whether the gestures phase
would execute any type of motion, as for instance circle, spiral, zigzag, or
straight, or whether only restricted set of movement types could be realized by
a particular gesture phase (feature pair [+-restricted]). “Flow of movement”
describes whether the movement could show accentuation or variation in its
velocity (feature pair [+-variable]).
The features of the categories just described are mutually exclusive, mean-
ing that if a phase, for instance, carries the feature [+movement] it cannot carry
the feature [−movement]. According to these two feature sets a preparation is
marked [+movement, +tension, −constant, +increasing, −variable, +restricted],
a hold is marked [−movement, +tension, +constant] (see Table 4 and 5).
In the second part of the paper, it has been shown how the sequential embed-
ding of the phases in specic linear successions results in changes of the f ormal
characteristics of particular phases. Accordingly, it is possible to distinguish
linear sequences, which systematically cause replacements of features in one
gesture phase as in the cases of “no execution of a preparation” or the “transi-
tion from a retraction to a rest position on the speakers body or to an action.”
In order to account for these phenomena, the format of writing phonological
rules (Chomsky and Halle 1968) was adapted. As the examination of these
examples has revealed, it is necessary to distinguish between the description
and analysis of gesture phases, i.e., a context-independent and a context-
sensitive description. In the former each gesture phase is depicted on its own
by means of the distinctive features and additional features. In the latter, the
inuence adjacent gesture phases may have on a phase is taken into account. It
has been pointed out that rst the units of the analysis need to be determined in
the context-independent description in order to account for variations in their
behavior, due to the replacement of features in a context-sensitive description.
5. Discussion
The articulatory description of gesture phases presented in this paper contrib-
utes to the ongoing research in gesture studies in several respects. It supple-
ments existing denitions of gesture phases focusing either primarily on their
relation to speech or their adjacency to preceding and following phases. By
offering an articulatory description divided into a context-independent and
context-sensitive section, it offers a clearer differentiation of the various levels
Bereitgestellt von | TU Chemnitz
Angemeldet
Heruntergeladen am | 26.03.15 10:39
Rethinking gesture phases: Articulatory features of gestural movement? 85
of analysis possible with regard to gesture phases. Considered as the rst ana-
lytical step in investigating gestures’ movement phases, it provides the founda-
tion for a further characterization. It spells out the articulatory characteristics
visible in the execution of gesture phases to which gesture analysts and coders
may attend when segmenting and coding the phases. The present proposal
therefore points out features of gesture phases that, if possibly supplemented
and enhanced by further characteristics and proposals, could be used for the
identication and coding of gesture phases. Accordingly, the paper contributes
to the ongoing discussion of how to code gesture phases and to the comparabil-
ity of gesture analyses. The application of a set of articulatory characteristics
provides for a possible consistency in the coding of gesture phases as well as
an agreement between different coders. Therefore, inter-rater reliability tests as
well as studies on the perception of gesture phases would be possible.
By and large the approach also contributes to the discussion on questions
regarding the (automatic) segmentation of gestures (e.g., Eisenstein and Ran-
dall 2004; Harling and Edwards 1997; Kettebekov and Sharma 2001; Sowa
2006) and their implementation in articial agents.
Furthermore, by concentrating on articulatory characteristics of gesture
phases, the current description may contribute to the discussion of “gestural-
ness” (Kendon 2004: 15), i.e., the features that gestures carry in order for them
to be perceived as such. Asking with Kendon, “what are the features that an
action must have for it to be treated as a gesture?” (Kendon 2004: 12), the
p aper may provide rst answers to the question as it explicates the features that
make gesture phases stand out against each other and possibly against other
types of actions.
By doing so, we did not aim at stating that gesture phases are akin to speech
phonemes. This might be read easily into our analysis as we exploit linguistic
analyses on sound structures in mind when we conceive of gesture phases as
possible bundles of distinctive features. Nonetheless, adopting a feature-based
perspective for the description of gesture phases must not necessarily imply the
assumption that the described gestural units resemble or are akin to units de-
scribed in speech. Rather the goal was to describe gesture phases by a means
of articulatory features. This was done to show that the phases can be charac-
terized separately by particular features and bundles of features, and are per-
ceivable and treatable as separate units of analysis. The phonological frame-
work thus served as an apparatus for description and did not aim at an equation
of the described units with linguistic units.
By providing a specication of the characteristics of gestures phases, the
paper offers the possibility of perceiving and treating them as a subject of in-
vestigation, which provides the opportunity to describe patterns and structures
on the level of the gesture phases themselves (see also Bressem in prep., Fricke
2008). Furthermore, it allows for a more ne-grained analysis of the i nterwoven
Bereitgestellt von | TU Chemnitz
Angemeldet
Heruntergeladen am | 26.03.15 10:39
86 J. Bressem and S. H. Ladewig
nature of linear and simultaneous aspects of speech and gestures, and for a
better integration of these two modalities into a single framework.
We will now make some concluding remarks on the theoretical framework
of the approach postulated in this paper. We approach the study of gestures
from a linguistic-semiotic viewpoint and especially a multimodal view on
grammar (Fricke 2007, 2008; Ladewig and Bressem under revision; Müller
1998, 2004; Müller, Fricke, Lausberg, Liebal in preparation). Accordingly,
gesture and speech are two distinct sign systems that exhibit common and spe-
cic semiotic properties that unite and distinguish the two modalities. First
studies conducted within a multimodal approach to grammar have shown ap-
parent common structural principles for language and gesture as, for instance,
proto-morphological structures in gestures (Müller 2004), the recursive em-
beddings of gesture phases within gesture units (Fricke 2008), and processes of
grammaticalizations working in recurrent gestural forms and gestures (Ladewig
and Bressem under revision). Furthermore, there seem to be points of struc-
tural integration of gestures into the syntax of speech, as gestures might func-
tion as attributes to nominal phrases (Fricke 2008) or are used in a noun or verb
position within a sentence (Ladewig in preparation). Starting with the assump-
tion that speech and gesture are two distinct systems while sharing common
properties, linguistic analyses of gestures aim at discovering commonalities
and overlapping characteristics as well as differences and specicities of the
two modalities. The objective is not to prove linguistic structures and proper-
ties of verbal language in gestures. Rather, the use of theoretical and method-
ological concepts from linguistics aims at extracting similarities and differ-
ences of the two sign systems by exploring their overlapping sets. Limits of
adaptability of linguistic concepts are therefore not conceived as obstacles in
proving gestures to be like a language, but are rather understood as necessary
limits in nding the characteristics of gestures and also of speech (see also
Birdwhistell 1970).
Suggesting a linguistic approach to the analysis of systems of communicative behavior
other than languages does not mean supposing that these other systems are languages.
It means, rather, adopting the level of analysis at which linguists operate when they ap-
proach speech and seeking for their mode of expression. Such an approach does assume,
of course, that when people interact they make use of a repertoire of behavioral forms
that they share with others and use in accordance with sets of shared rules. This does not
presuppose the nature of these forms, however, nor does it presuppose the nature of
their rules of patterning. These may be quite different from those that may be found in
language. (Kendon 1972b: 443– 444)
The linguistic approach thereby serves as a theoretical building block from
which elements are selectively taken and carefully adopted in the analysis of
gestures. This adaptation is always guided by the cautiousness of not imposing
Bereitgestellt von | TU Chemnitz
Angemeldet
Heruntergeladen am | 26.03.15 10:39
Rethinking gesture phases: Articulatory features of gestural movement? 87
linguistic structures onto gestures, which they themselves do not exhibit.
Nevertheless, speaking with Pike (1967), “verbal and nonverbal activity is a
unied whole, and theory and methodology should be organized to treat it
as such” (Pike 1967: 26).
In doing so, a linguistic approach is a companion to present foci, such as
psychological or interactional approaches, by expanding the elds of investi-
gations and approaches in gesture studies and thereby contributing to a more
thorough understanding of the medium “gesture” itself as well as the relation
of speech and gesture. It allows for a different point of view on phenomena
observable in gestures and a perspective of description so far missing in ges-
tures studies.
Notes
* This paper was written within the project “Towards a grammar of gesture: Evolution, brain,
and linguistic structures” which is funded by the Volkswagen foundation (see http://www.
togog.org). First ideas of this gesture phase description were presented in a talk given in the
research colloquium “Multimodality” at the European-University Viadrina, Frankfurt (Oder)
on 18 May, 2007. The revised and nal version, on which the paper is based, was presented
at the MGA I on October 11th 2007. We thank Ellen Fricke for helpful comments on an earlier
version of this paper and the members of the togog-group as well as the participants of the
MGA I Workshop for their suggestions.
1. With the term “gesture,” we refer to instances of strokes as well as strokes in concjunction
with holds, i.e., gesture phrases in the sense of Kendon (1980, 2004).
2. Contrary to his proposal in 1980, Kendon excludes the phase retraction from the “gesture
phrase” and denes it as being made up of the phases “preparation,” “stroke,” and “post-
stroke hold” (Kendon, 2004: 112). Furthermore, Kendon incorporates the notion of the “post-
stroke hold” and steps back from the notion of the “partial retraction” in his latest work.
3. First considerations on this topic go back to Kita (1990), where he puts forward the idea of
functionally different types of holds.
4. ELAN (Eudico Linguistic Annotator) has been developed at the Max Plank Institute for
Psycholinguistics in Nijmegen (see http://www.lat-mpi.eu/tools/elan/ for further i nformation).
5. Unlike Fricke (2008), we do not include the parameter gravitation to describe the articula-
tory features of the gesture phases. According to Fricke, the amount of muscular tension “is
measured by which and how many muscles of the hand and arm are executed antagonisti-
cally to gravitational force” (Fricke 2008: 158, our translation). Accordingly, the impact of
gravitational force remains constant during the execution of a stroke. The antagonistic
m uscle movements produce constant tension and remain constantly articulated (Fricke 2008:
194). However, when using Fricke’s proposed features, phenomena such as accentuated
strokes (see Section 3) could not be described as gravitational force does not remain constant
in these cases but changes. Concluding from this, one could reason that Fricke also uses
movement and tension as the driving features to determine the gesture phases and not, as
proposed, gravitational force.
6. The terms “pronation” and “supination” have been introduced into the description of sign
language by Stokoe (1972) and into the study of gestures by Ott (1892), Efron (1972 [1941]),
Kendon (2004), Sparhawk (1978), Sager (2005) or Weinrich (1992).
Bereitgestellt von | TU Chemnitz
Angemeldet
Heruntergeladen am | 26.03.15 10:39
88 J. Bressem and S. H. Ladewig
7. Attempts in gesture research at capturing the movement qualities of strokes usually refer to
the “effort-shape” approach as put forward by Laban and Lawrence for the notation of dance.
We are familiar with the notation and the effort and shape characteristics as well as ap-
proaches trying to characterize gestures based on them (cf. Martell 2005; Sowa 2006; Zhao
2001), but, nevertheless, refrain from using “effort-shape” terminology for the description of
gesture phases. Speaking with Martell, we “do not yet accept their irreducibility. For present
purposes, we see Effort and Shape as describing equivalence classes of movements. Our in-
terest is in understanding the low-level physical aspects that make up these classes of move-
ments” (Martell 2005: 30).
8. The terms “body-focused” and “object-focused” movements have been introduced by Freed-
man (1972, 1977). He denes object-focused movements as “representing behavior and
body-focused movements as attention sustaining behavior” (Freedman 1977: 118). The for-
mer are executed in front of the body, depict the content of a message or the “form of what is
being represented” (Freedman 1977: 121). The latter refer to movements of self-stimulation
and fulll regulating functions, as they appear to be effective adaptations “to stress, allowing
for the sustained processing of thought” (Freedman 1977: 124).
9. By accentuation of movement, we understand that parts of the motion are stressed such that
the movement is carried out with more force. This rise in force leads to an increase in the
intensity of the movement execution. Similarly to the accent in the spoken language, in
which the accent is used to stress particular segments of speech such as syllables, for instance
(see for example Pompino-Marschall 1995), the accentuation of gestural movements is used
to stress a particular gestural segment (see also Bressem in preparation for a more detailed
discussion of accentuation patterns in gestural movement patterns).
10. We are grateful to Mathias Roloff for providing the drawings (http://www.mathiasroloff.de).
References
Arendson, Jeroen, Andrea J. van Doorn & Huib de Ridder. 2007. When and how well do people
see the onset of gestures. Gesture 7(3). 305–342.
Birdwhistell, Ray L. 1970. Kinesics and context. Philadelphia: University of Pennsylvania Press.
Bohle, Ulrike. 2007. Das Wort ergreifen — das Wort übergeben. Explorative Studie zur Rolle re-
debegleitender Gesten in der Organisation des Sprecherwechsels. Berlin: Weidler Verlag.
Bressem, Jana. 2008. Characterizing gestural form features — Suggestions for a form based nota-
tional system of coverbal gestures. http://www.janabressem.de/ Downloads/ Bressem_notating
%20gestures.pdf (accessed 1 December 2010)
Bressem, Jana (in preparation). Repetitions in gesture: Structures and cognitive aspects. Frankfurt:
European-University Viadrina dissertation.
Chafai, Nicolas, Catherine Pelachaud & Pelé Danielle. 2006. Analysis of gesture expressivity
modulations from cartoons animations. Paper presented at the LREC 2006 workshop on “multi-
modal corpora,” Genova, May.
Chomsky, Noam. 1993. A minimalist program for linguistic theory. In K. Hale & S. J. Keyser
(eds.), The view from building 20, 1–52. Cambridge: MIT Press.
Chomky, Noam & Morris Halle. 1968. The sound pattern of English. New York: Harper and Row.
Condon, William S. & William D. Ogston. 1967. A segmentation of behavior. Journal of Psychi-
atric Research 51. 221–235.
Duncan, S. D. n. d. Coding “manual.” http://mcneilllab.uchicago.edu/pdfs/Coding_Manual.pdf
(accessed 29 November 2010).
Efron, D. 1972 (1941). Gesture, race, and culture. Paris & The Hague: Mouton.
Bereitgestellt von | TU Chemnitz
Angemeldet
Heruntergeladen am | 26.03.15 10:39
Rethinking gesture phases: Articulatory features of gestural movement? 89
Eisenstein, Jacob & Davis Randall. 2004. Visual and linguistic information in gesture classica-
tion. In Proceedings of the sixth international conference on multimodal interfaces, 113–120.
New York: ACM.
Ekman, Paul & Wallace Friesen. 1969. The repertoire of nonverbal behavior: Categories, origins,
usage, and coding. Semiotica 1. 49–98.
Ekman, Paul & Wallace Friesen. 1974. Detecting deception from the body or face. Journal of
Personality and Social Psychology 29. 288–298.
Ekman, Paul & Wallace Friesen. 1977. Hand movements. Journal of Communication 22. 353–374.
Freedman, Norbert. 1972. The analysis of movement behavior during the clinical interview. In
Aron Wolfe Siegman & Benjamin Pope (eds.), Studies in dyadic communication, 153–175. New
York: Pergamon.
Freedman, Norbert. 1977. Hands, words, and mind: On the structuralization of body movements
during discourse and the capacity for verbal representation. In Norbert Freedman & Stanley
Grand (eds.), Communicative structures and psychic structures, 109–132. New York: Plenum.
Fricke, Ellen. 2007. Origo, Geste und Raum: Lokaldeixis im Deutschen. Berlin & New York:
Mouton de Gruyter.
Fricke, Ellen. 2008. Grundlagen einer multimodalen Grammatik des Deutschen: Syntaktische
Strukturen und Funktionen. Unpublished Habilitationsschrift, Europa-Universität Viadrina
Frankfurt (Oder).
Gullberg, Marianne & Kenneth Holmquist. 2006. What speakers do and what addresses look at:
Visual attention to gestures in human interaction live and on video. Pragmatics & Cognition
14(1). 53–82.
Hall, Alan T. 2000. Phonologie: Eine Einführung. Berlin & New York: Walter de Gruyter.
Harling, Phillip & Alistair D. N. Edwards. 1997. Hand tension as a gesture segmentation cue. In
Progress in gestural interaction: Proceedings of gesture workshop 1996, University of York, 19
March, 1996, 75–88. London: Springer Verlag.
Kahol, Kanav, Priyamvada Tripathi & Sethuraman Panchanathan. 2004. Automated gesture
s egmentation from dance sequences. In Proceedings of the IEEE international conference
on automatic face and gesture recognition. http://www.public.asu.edu/~kkahol/publications/
kanavfandgesture2004.pdf (accessed 29 November 2010).
Kendon, Adam. 1972a. Some relationship between body motion and speech. In Aaron Siegman &
Benjamin Pope (eds.), Studies in dyadic communication, 177–216. Elmsford: Pergamon Press.
Kendon, Adam. 1972b. Review of kinesics and context. American Journal of Psychology 85.
441– 456.
Kendon, Adam. 1980. Gesture and speech: Two aspects of the process of utterance. In Mary R.
Key (ed.), Nonverbal communication and language, 207–227. The Hague: Mouton.
Kendon, Adam. 1996. An agenda for gesture studies. Semiotic Review of Books 7(3). 8–12.
Kendon, Adam. 2004. Gesture: Visible action as utterance. Cambridge: Cambridge University
Press.
Kettebekov, Sanshzar & Rajeev Sharma. 2001. Toward natural gesture/speech control of a large
display. In R. Little & Laurence Nigay (eds.), Engineering for human-computer interaction,
221–234. Heidelberg: Springer Verlag.
Kita, Sotaro. 1990. The temporal relationship between gesture and speech: A study of Japanese-
English bilinguals. Chicago: University of Chicago master’s thesis.
Kita, Sotaro, Ingeborg van Gijn & Harry van der Hulst. 1998. Movement phases in signs and co
speech gestures and their transcription by human encoders. In Ipke Wachsmuth & Martin
Fröhlich (eds.), Gesture and sign language in human-computer interaction, 23–35. Berlin:
Springer.
Ladewig, Silva H. 2007. The cyclic gesture and its variants — a new gesture family? Unpublished
manuscript.
Bereitgestellt von | TU Chemnitz
Angemeldet
Heruntergeladen am | 26.03.15 10:39
90 J. Bressem and S. H. Ladewig
Ladewig, Silva H. 2010. Beschreiben, auffordern und suchen — Varianten einer rekurrenten
Geste. Zeitschrift für Sprache und Literatur 41(1). 89–111.
Ladewig, Silva H. in press. Putting a recurrent gesture on a cognitive basis. CogniTextes.
Ladewig, Silva H. in preparation. Syntactic and semantic integration of gestures into speech:
Structural, cognitive and conceptual aspects. Frankfurt: European-University Viadrina
d issertation.
Ladewig, Silva H. & Jana Bressem. under revision. Discovering structures in gestures based on the
four parameters of sign language.
Latoschik, Marc Erich. 2000. Multimodale Interaktion in virtueller Realität am Beispiel der virtu-
ellen Konstruktion. Bielefeld: Universität Bielefeld dissertation.
Loehr, Dan. 2006. Gesture and intonation. Washington: Georgetown University dissertation.
Martell, Craig. 2005. FORM: An experiment in the annotation of the kinematics of gesture. Phila-
delphia: University of Pennsylvania Dissertation.
Martell, Craig & Joshua Kroll. 2007. Corpus-based gesture analysis: An extension of the FORM
dataset for the automatic detection of phases in gesture. International Journal of Semantic Com-
puting 1(4). 521–536.
McClave, Evelyn. 1991. Intonation and gesture. Washington: Georgetown University dissertation.
McNeill, David. 1992. Hand and mind. Chicago: Chicago University Press.
McNeill, David. 2005. Gesture and thought. Chicago: Chicago University Press.
McNeill, David. in press. Gesture and thought. In C. Patrick Hogan (ed.), Cambridge encyclopedia
of the language sciences.
Menzerath, Paul & Antonio de Lacerda. 1933. Koartikulation, Steuerung und Lautabgrenzung.
Berlin & Bonn: Ferd. Dümmlers Verlag.
Müller, Cornelia. 1994. Cómo se llama . . . ? Kommunikative Funktionen des Gestikulierens in
Wortsuchen. In Peter-Paul König & Helmut Wiegers (eds.), Satz — Text — Dikurs, 71–80.
Tübingen: Niemeyer.
Müller, Cornelia. 1998. Redebegleitende Gesten. Kulturgeschichte — Theorie — Sprachvergleich.
Berlin: Berlin Verlag.
Müller, Cornelia. 2004. Forms and uses of the palm up open hand: A case of a gesture family? In
Cornelia Müller & Roland Posner (eds.), Semantics and pragmatics of everyday gestures: Pro-
ceedings of the Berlin conference April 1998, 233–256. Berlin: Berlin Verlag.
Müller, Cornelia, Lausberg Hedda, Liebal Katja & Ellen Fricke (in preparation). Gestural modes
of representation or how hands turn into gestures: Semiotic structures, neurological foundations,
and evolutionary implications.
Nobe, Shuichi. 2000. Where do most spontaneous representational gestures actually occur with
respect to speech? In David McNeill (ed.), Language and gesture, 186 –198. Cambridge: Cam-
bridge University Press.
Ott, Edward Amherst. 1892. How to gesture. New York: Hinds & Noble.
Parill, Fey. 2001. Hand to mouth: Linking spontaneous gesture and aspect. Unpublished Berkeley:
University of Berkeley dissertation.
Pike, K. 1967. Language in relation to a unied theory of the structure of human behavior. The
Hague: Mouton.
Pompino-Marschall, Bernd. 1995. Einführung in die Phonetik. Berlin & New York: Walter de
Gruyter.
Queck, Francis, David McNeill, Robert Bryll, Susan Duncan, Xing-Feng Ma, Cemil Kirbas, Karl
E. McCullough & Rashid Ansari. 2002. Multimodal human discourse: Gesture and speech.
ACM Transactions on Computer-Human Interaction 9(3). 171–193.
Sager, Sven F. 2005. Ein System zur Beschreibung von Gestik. Osnabrücker Beiträge zur Sprach-
theorie (OBST) 70. 19– 47.
Bereitgestellt von | TU Chemnitz
Angemeldet
Heruntergeladen am | 26.03.15 10:39
Rethinking gesture phases: Articulatory features of gestural movement? 91
Seyfeddinipur, Mandana. 2006. Disuency: Interrupting speech and gesture (MPI Series in Psy-
cholinguistics 39). Nimegen: University of Nijmegen.
Sowa, Timo. 2006. Understanding coverbal iconic gestures in object shape descriptions. Berlin:
Akademische Verlagsgesellschaft Aka GmbH. Berlin.
Sparhawk, Carol. 1978. Contrastive-identicational features of Persian gesture. Semiotica 24(1/2).
49–86.
Stokoe, W. C. 1972. Semiotics and human sign language. The Hague & Paris: Mouton.
Tag, Susanne. in preparation. Retractions, holds and rest positions: Three gesture phases under
examination.
Ternes, Elmar. 1999. Einführung in die Phonologie. Darmstadt: Wissenschaft-liche Buchgesell-
schaft.
Trubetzkoy, Nikolai. S. 1958. Grundzüge der Phonologie. Göttingen: Vandenhoeck & Ruprecht.
Webb, Rebecca. 1996. Linguistic features of metaphoric gestures. Rochester: University of Roch-
ester dissertation.
Wilson, Andrew D., Aaron Bobick & Justine Cassell. 1997. Temporal classication of natural
gesture and application to video coding. IEEE computer society conference on computer vision
and pattern recognition, 17–19. 948–954.
Weinrich, Lotte. 1992. Verbale und nonverbale Strategien in Fernsehgesprächen. Eine explorative
Studie. Tübingen: Niemeyer.
Yassinik, Yelena, Margret Renwick & Stefanie Shattuck-Hufnagel. 2004. The timing of speech
accompanying gestures with respect to prosody. http://www.rle.mit.edu/soundtosense/
c onference/pdfs/fulltext/ Friday%20Posters/ FA-Yasinnik-STS-MAC.pdf (29 November 2010)
Zhao, Liwei. 2001. Synthesis and acquisition of Laban movement analysis qualitative parameters
for communicative gestures. Philadelphia: University of Pennsylvania dissertation.
Jana Bressem ( b. 1979) is a research assistant at European University Viadrina <research@
janabressem.de>. Her research interests include descriptions and analyses of gestures from a
l inguistic-semiotic point of view both in human and non-human primates.
Silva H. Ladewig ( b. 1978) is a research assistant at European University Viadrina <sladewig@
cgest.de>. Her research interests include descriptions and analyses of gestures from a linguistic-
semiotic as well as a cognitive-linguistic view point.
Bereitgestellt von | TU Chemnitz
Angemeldet
Heruntergeladen am | 26.03.15 10:39
